Data set information: https://archive.ics.uci.edu/ml/datasets/Fertility
Must compile with C++11
Example of input and output:
$ ./a.out
Welcome to Luis Lopez's Feature Selection Algorithm.
Type in the name of the file to test: 72.txt
Type the number of the algorithm you want to run.
- Forward Selection
- Backward Elimination
- Luis's Special Algorithm
1 This dataset has 10 features (not including the class attribute), with 100 instances.
Please wait while I normalize the data... Done!
Beginning search
Using features(s) {1} accuracy is 69%
Using features(s) {2} accuracy is 74%
Using features(s) {3} accuracy is 74%
Using features(s) {4} accuracy is 69%
Using features(s) {5} accuracy is 65%
Using features(s) {6} accuracy is 82%
Using features(s) {7} accuracy is 65%
Using features(s) {8} accuracy is 75%
Using features(s) {9} accuracy is 73%
Using features(s) {10} accuracy is 68%
Feature set {6} was best, accuracy is 82%
Using features(s) {6,1} accuracy is 80%
Using features(s) {6,2} accuracy is 98%
Using features(s) {6,3} accuracy is 88%
Using features(s) {6,4} accuracy is 85%
Using features(s) {6,5} accuracy is 84%
Using features(s) {6,7} accuracy is 86%
Using features(s) {6,8} accuracy is 82%
Using features(s) {6,9} accuracy is 82%
Using features(s) {6,10} accuracy is 87%
Feature set {6,2} was best, accuracy is 98%
Using features(s) {6,2,1} accuracy is 97%
Using features(s) {6,2,3} accuracy is 92%
Using features(s) {6,2,4} accuracy is 95%
Using features(s) {6,2,5} accuracy is 91%
Using features(s) {6,2,7} accuracy is 94%
Using features(s) {6,2,8} accuracy is 96%
Using features(s) {6,2,9} accuracy is 97%
Using features(s) {6,2,10} accuracy is 95%
(Warning, Accuracy has decreased! Continuing search in case of local maxima)
Feature set {6,2,1} was best, accuracy is 97%
Using features(s) {6,2,1,3} accuracy is 88%
Using features(s) {6,2,1,4} accuracy is 87%
Using features(s) {6,2,1,5} accuracy is 90%
Using features(s) {6,2,1,7} accuracy is 88%
Using features(s) {6,2,1,8} accuracy is 91%
Using features(s) {6,2,1,9} accuracy is 95%
Using features(s) {6,2,1,10} accuracy is 89%
(Warning, Accuracy has decreased! Continuing search in case of local maxima)
Feature set {6,2,1,9} was best, accuracy is 95%
Using features(s) {6,2,1,9,3} accuracy is 93%
Using features(s) {6,2,1,9,4} accuracy is 89%
Using features(s) {6,2,1,9,5} accuracy is 86%
Using features(s) {6,2,1,9,7} accuracy is 88%
Using features(s) {6,2,1,9,8} accuracy is 89%
Using features(s) {6,2,1,9,10} accuracy is 94%
(Warning, Accuracy has decreased! Continuing search in case of local maxima)
Feature set {6,2,1,9,10} was best, accuracy is 94%
Using features(s) {6,2,1,9,10,3} accuracy is 86%
Using features(s) {6,2,1,9,10,4} accuracy is 88%
Using features(s) {6,2,1,9,10,5} accuracy is 87%
Using features(s) {6,2,1,9,10,7} accuracy is 84%
Using features(s) {6,2,1,9,10,8} accuracy is 83%
(Warning, Accuracy has decreased! Continuing search in case of local maxima)
Feature set {6,2,1,9,10,4} was best, accuracy is 88%
Using features(s) {6,2,1,9,10,4,3} accuracy is 84%
Using features(s) {6,2,1,9,10,4,5} accuracy is 85%
Using features(s) {6,2,1,9,10,4,7} accuracy is 82%
Using features(s) {6,2,1,9,10,4,8} accuracy is 87%
(Warning, Accuracy has decreased! Continuing search in case of local maxima)
Feature set {6,2,1,9,10,4,8} was best, accuracy is 87%
Using features(s) {6,2,1,9,10,4,8,3} accuracy is 76%
Using features(s) {6,2,1,9,10,4,8,5} accuracy is 87%
Using features(s) {6,2,1,9,10,4,8,7} accuracy is 78%
(Warning, Accuracy has decreased! Continuing search in case of local maxima)
Feature set {6,2,1,9,10,4,8,5} was best, accuracy is 87%
Using features(s) {6,2,1,9,10,4,8,5,3} accuracy is 74%
Using features(s) {6,2,1,9,10,4,8,5,7} accuracy is 79%
(Warning, Accuracy has decreased! Continuing search in case of local maxima)
Feature set {6,2,1,9,10,4,8,5,7} was best, accuracy is 79%
Using features(s) {6,2,1,9,10,4,8,5,7,3} accuracy is 73%
(Warning, Accuracy has decreased! Continuing search in case of local maxima)
Feature set {6,2,1,9,10,4,8,5,7,3} was best, accuracy is 73%
Finished search! The best feature subset is {6,2} which has an accuracy of 98%